This invention relates to the processing of colour television signals and is particularly concerned with the separation within an encoded television signal, of luminance and chrominance information.
An encoded television signal is made up of a luminance signal (Y), and colour difference signals (U, V) which define hue and saturation. In both NTSC and PAL television systems, the two colour difference signals are each used to amplitude modulate a colour sub-carrier using quadrature modulation. The resulting signal is referred to as the chrominance signal. The colour sub-carrier frequency necessarily lies within the bandwidth of the luminance signal and it is the function of a decoder circuit to decode the luminance and chrominance information avoiding, so far as possible, cross-talk.
Decoder circuits are used in a variety of applications. The decoder circuit within a television receiver is perhaps the simplest application since the decoded signal is used directly for the television display. In other applications, a decoded signal undergoes further processing and may be subsequently re-encoded. In applications of this sort, the demands placed upon the decoder circuit are of course higher.
Different television standards, principally PAL and NTSC, employ different techniques for the encoding of luminance and chrominance information. In PAL, for example, the sub-carrier used for one of the colour difference signals is inverted on alternate lines at the television picture. The sub-carrier frequency is related to the television line scanning frequency but is off-set from an integral multiple of line frequency by one quarter of line frequency. In NTSC, the off-set is one half of the line frequency. Because of these, and other differences, separate PAL and NTSC encoders and decoders are developed. This specification will concentrate on PAL but it will be recognised that there are parallels throughout in the encoding and decoding of PAL, NTSC andxe2x80x94indeedxe2x80x94other standards.
A rudimentary PAL decoder provides a luminance signal by the use of a notch or band stop filter designed to suppress the colour sub-carrier frequency. Similarly, a band pass filter can provide a chrominance signal. Chrominance information is centered at 4.43 MHz but may be spread over the range 3.3 to 5.5 MHz. It will be apparent that the simple use of a notch filter will attenuate luminance information at or near 4.43 MHz, resulting in a loss of horizontal resolution. Also, chrominance information away from 4.43 MHz may be attenuated only partially and will thus be misinterpreted in subsequent circuitry as luminance information. Nonetheless, the use of a notch filter has the merit of simplicity and cheapness.
Because there is a well defined relationship between the colour difference signals on alternating lines, it is possible to derive a filter arrangement, typically utilising two line delays, to produce a filtered chrominance signal. Such circuits are commonly referred to as comb filters and a wide variety of proposals have been made for comb filters suited to varying applications.
A comb filter can in theory be designed to remove all chrominance information in the special case where there is no change, line by line, in the luminance or chrominance signals. In most practical situations, however, there will be small amounts of cross-talk. Depending upon the environment in which the encoder is used, the design can be optimised to reduce the perceived effects of these errors.
In one prior proposal, a measurement is made of the degree to which the luminance and chrominance information departs from the assumed ideal of constancy and, where a threshold amount of difference is exceeded, the comb filter is disabled and an alternative form of decoderxe2x80x94such as a notch filterxe2x80x94used in its place.
The described line comb filter has, in certain prior proposals, been extended into two and sometimes three dimensions. That is to say appropriate delays can be introduced to permit comparisons between pixels in neighbouring lines, between pixels in the same line and between pixels in succeeding fields. The degree of weighting to be applied to the samples in a one, two or three dimensional filter will be selected to optimise performance. Indeed, in one prior proposal, signals derived from a picture analyser are used dynamically to vary the coefficients in a two dimensional filter.
An inherent difficulty in the use of comb filters, of whatever dimension, is that the phase of signals at various points within the decoder must be carefully controlled. It will be recognised that a very slight phase difference in the subtraction of a composite signal and a comb filtered chrominance signal can lead to significant decoding errors. For this reason, whilst comb filters are able to provide significant improvements in decoding over band stop and band pass filters, there are significant penalties in the complexity of the circuitry and in costs.
It is one object of this invention to provide an improved decoder which offers an improvement in accuracy over band stop and band pass filtering without placing unduly high demands on phase matching.
There are close parallels between encoding and decoding and numerous other processes relating to the separation of luminance and chrominance. This invention is not therefore restricted to decoding.
It is a further object of this invention to provide an improved encoder capable, in specified applications, of providing desired encoder characteristics in a more efficient manner than hitherto.
It is still a further object of this invention to provide an improved luminance filter for use in a variety of applications.
Accordingly, in one aspect the present invention consists in a colour television signal processor for use in the separation of chrominance and luminance signals, comprising an input terminal and an output terminal, first filter means connected with the input terminal for producing a first filtered signal, gate means having a control terminal and adapted for selectively passing the first filtered signal to the output terminal, and second filter means connected with the input terminal for delivering a second filtered signal to the control terminal of the gate means.
Advantageously, the first filter means comprises a band pass filter at the colour sub-carrier frequency.
Preferably, the second filter means comprises a comb filter arrangement.
Suitably, the gate means comprises a first gate having a first control terminal and being connected to receive a positive going component of the first filtered signal, and a second gate having a second control terminal and being connected to receive a negative going component of the first filter signal, the first and second control terminals receiving positive and negative going components, respectively, of the second filtered signal.
In a further aspect the present invention consists in a colour television processor adapted to produce luminance and chrominance signals from a composite television signal, comprising:
a chroma filter circuit adapted to produce from the composite signal a chrominance signal;
signal processing means for generating from said chrominance signal a detected chrominance signal;
band filter means adapted to produce from the composite signal a band filtered chrominance signal;
gate means receiving the band filtered chrominance signal and serving under the control of the detected chrominance signal to pass a gated chrominance signal; and
subtraction means for generating a luminance signal through subtraction of the composite television signal and gated chrominance signal.
Suitably, the chroma filter circuit comprises one or more delays and is adapted to produce a comb filtered chrominance signal.
In the decoder according to a preferred form of the present invention a luminance signal is generated by the subtraction from the composite signal of a chrominance signal which is band filleted but which has not been subject to comb filtering. There is accordingly a straightforward and readily maintained exact phase relationship between the signals which are to be subtracted. To deal with the problem that the band filtered signal will often contain spurious luminance signals at or around the colour sub-carrier frequency, the band filtered signal, prior to subtraction from the composite signal, is gated in accordance with the detected chrominance signal. This latter signal is suitably derived from a comb filtering process and is thus capable of distinguishing such spurious luminance information. Since, however, the comb filtered chrominance signal is used to control a gate and is not directly subtracted from the composite signal, slight phase errors in the comb filter are insignificant. Indeed, the output of the comb filter circuit may undergo rectification and/or envelope detection to produce the detected chrominance signal for use in controlling the gate.